Decision Frameworks

1. Introduction to Decision Frameworks

A decision framework is a structured approach to making choices. Frameworks reduce cognitive load, ensure consistency, and help teams align on how decisions get made—especially when stakes are high or multiple stakeholders are involved.

flowchart LR
    R["R — Recommend\nProposes solution"] --> I["I — Input\nProvides expertise"]
    I --> A["A — Agree\nMust approve\n(veto power)"]
    A --> D["D — Decide\nMakes final call"]
    D --> P["P — Perform\nExecutes decision"]
                        

Why Use Frameworks?

• Clarity: Everyone knows the process and their role
• Speed: Reduces debate about how to decide
• Accountability: Clear ownership of decisions
• Quality: Systematic consideration of factors
• Scalability: Works across teams and decision types

Choosing the Right Framework

2. RAPID Framework

RAPID (by Bain & Company) clarifies who is involved in a decision and what role each person plays. It's especially useful for cross-functional decisions where accountability is unclear.

Understanding RAPID Roles

R - RECOMMEND:   Proposes options, gathers data, drives analysis
A - AGREE:       Must agree before decision proceeds (veto power)
P - PERFORM:     Executes once decision is made
I - INPUT:       Provides info/expertise (no decision authority)
D - DECIDE:      Makes the final call (only ONE person)

Key Rule: Only ONE person has the "D" (Decide) role!

Implementation Guide

1. Define the decision clearly and specifically
2. Identify stakeholders who need to be involved
3. Assign exactly one D (the most common mistake is multiple Ds)
4. Minimize A roles (veto power should be rare)
5. Communicate the framework to all parties
6. Document the decision and rationale

3. RACI Matrix

RACI is similar to RAPID but focused on project/task responsibility rather than single decisions.

RACI Components

• R - Responsible: Does the work
• A - Accountable: Ultimately answerable (only one per task)
• C - Consulted: Provides input (two-way communication)
• I - Informed: Kept in the loop (one-way communication)

RACI vs. RAPID

4. Decision Matrices

Decision matrices help compare options across multiple criteria systematically.

Weighted Scoring Models

Steps:

1. List your options (rows)
2. Define criteria (columns)
3. Assign weights to criteria (must sum to 100%)
4. Score each option (e.g., 1-5 scale)
5. Calculate weighted scores

Pugh Matrix

Compares alternatives against a baseline using +, -, or S (same):

• Select a baseline option (often the current state or leading option)
• Compare each alternative to baseline on each criterion
• Count pluses and minuses

Advantage: Simpler than weighted scoring; good for early screening.

5. Decision Trees

Decision trees map sequential decisions and their outcomes, incorporating uncertainty.

Tree Construction

□ = Decision Node    ○ = Chance Node    ▢ = Outcome

□ Launch Now?
├─ YES
│   ○ Market Response
│   ├─ Strong (60%) → ▢ +$5M profit
│   └─ Weak (40%)   → ▢ -$1M loss
│
└─ NO (Delay 6 months)
    ○ Competitor Moves
    ├─ No entry (70%) → ▢ +$3M profit
    └─ Entry (30%)    → ▢ +$1M profit

graph TD
    D{"Decision:\nLaunch Product?"} -->|"Launch"| C1{"Market\nResponse"}
    D -->|"Don't Launch"| ZERO["$0\n(Status Quo)"]
    
    C1 -->|"Success (60%)"| WIN["Revenue: $5M\nProfit: $3M"]
    C1 -->|"Moderate (25%)"| MED["Revenue: $2M\nProfit: $500K"]
    C1 -->|"Failure (15%)"| LOSE["Revenue: $500K\nLoss: -$1M"]
    
    WIN -.-> EV["EMV = (0.6 × $3M)\n+ (0.25 × $500K)\n+ (0.15 × -$1M)\n= $1.775M"]
                        

Expected Value Analysis

Calculate expected value (EV) for each path:

• Launch Now: (0.60 × $5M) + (0.40 × -$1M) = $3M - $0.4M = $2.6M
• Delay: (0.70 × $3M) + (0.30 × $1M) = $2.1M + $0.3M = $2.4M

Decision: Launch now has higher expected value ($2.6M vs $2.4M).

6. Cognitive Biases in Decision Making

Even with frameworks, human judgment is susceptible to systematic biases.

Common Decision Biases

Debiasing Strategies

• Pre-mortem: "Assume this failed—what went wrong?" Surfaces risks
• Devil's advocate: Assign someone to argue against the decision
• Outside view: Ask "What happens to similar projects?" (base rates)
• Red team/Blue team: Separate teams argue for and against
• Blind analysis: Remove identifying info when evaluating options
• Decision journals: Record predictions and revisit to calibrate

7. Conclusion & Next Steps

You've now covered the key concepts in this section of data-driven decision making. Here's a summary of what you've learned:

In the next article, we'll cover Data Collection & Quality Management—the foundation for all reliable analytics and decision-making.